INDEX

Explanations

phrases or sentences emphasizing qualities, conditions, or comparisons

oai_token-act-pair · gpt-3.5-turbo

negating phrases that emphasize limits or exceptions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ibaba

-0.89

etsk

-0.67

 separat

-0.63

 Pengu

-0.63

ovsky

-0.60

ashtra

-0.59

abi

-0.57

 havoc

-0.57

ario

-0.57

 Parenthood

-0.56

POSITIVE LOGITS

 means

0.81

 virtue

0.79

 leaps

0.74

uu

0.72

 Means

0.71

umbers

0.69

dB

0.67

 margins

0.67

proxy

0.65

products

0.65

Activations Density 0.228%

phrases or sentences emphasizing qualities, conditions, or comparisons

negating phrases that emphasize limits or exceptions

No Comments

No Known Activations